{"id":"https://openalex.org/W7130401414","doi":"https://doi.org/10.48550/arxiv.2602.15720","title":"ToaSt: Token Channel Selection and Structured Pruning for Efficient ViT","display_name":"ToaSt: Token Channel Selection and Structured Pruning for Efficient ViT","publication_year":2026,"publication_date":"2026-02-17","ids":{"openalex":"https://openalex.org/W7130401414","doi":"https://doi.org/10.48550/arxiv.2602.15720"},"language":null,"primary_location":{"id":"pmh:doi:10.48550/arxiv.2602.15720","is_oa":true,"landing_page_url":null,"pdf_url":null,"source":{"id":"https://openalex.org/S4406922384","display_name":"Open MIND","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"repository"},"license":"publisher-specific-oa","license_id":"https://openalex.org/licenses/publisher-specific-oa","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Article"},"type":"preprint","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":null,"any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5126290795","display_name":"Hyunchan Moon","orcid":null},"institutions":[],"countries":[],"is_corresponding":true,"raw_author_name":"Moon, Hyunchan","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5034556789","display_name":"Cheonjun Park","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Park, Cheonjun","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5126347310","display_name":"Steven L. Waslander","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Waslander, Steven L.","raw_affiliation_strings":[],"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":3,"corresponding_author_ids":["https://openalex.org/A5126290795"],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":null,"last_page":null},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10036","display_name":"Advanced Neural Network Applications","score":0.5598999857902527,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},"topics":[{"id":"https://openalex.org/T10036","display_name":"Advanced Neural Network Applications","score":0.5598999857902527,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T10502","display_name":"Advanced Memory and Neural Computing","score":0.11140000075101852,"subfield":{"id":"https://openalex.org/subfields/2208","display_name":"Electrical and Electronic Engineering"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T11019","display_name":"Image Enhancement Techniques","score":0.07909999787807465,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/trimming","display_name":"Trimming","score":0.6205999851226807},{"id":"https://openalex.org/keywords/pruning","display_name":"Pruning","score":0.5277000069618225},{"id":"https://openalex.org/keywords/selection","display_name":"Selection (genetic algorithm)","score":0.5148000121116638},{"id":"https://openalex.org/keywords/security-token","display_name":"Security token","score":0.45969998836517334},{"id":"https://openalex.org/keywords/channel","display_name":"Channel (broadcasting)","score":0.4332999885082245},{"id":"https://openalex.org/keywords/code","display_name":"Code (set theory)","score":0.3361999988555908},{"id":"https://openalex.org/keywords/artificial-neural-network","display_name":"Artificial neural network","score":0.30149999260902405},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.30090001225471497}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7450000047683716},{"id":"https://openalex.org/C56951928","wikidata":"https://www.wikidata.org/wiki/Q3539213","display_name":"Trimming","level":2,"score":0.6205999851226807},{"id":"https://openalex.org/C108010975","wikidata":"https://www.wikidata.org/wiki/Q500094","display_name":"Pruning","level":2,"score":0.5277000069618225},{"id":"https://openalex.org/C81917197","wikidata":"https://www.wikidata.org/wiki/Q628760","display_name":"Selection (genetic algorithm)","level":2,"score":0.5148000121116638},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4805999994277954},{"id":"https://openalex.org/C48145219","wikidata":"https://www.wikidata.org/wiki/Q1335365","display_name":"Security token","level":2,"score":0.45969998836517334},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.44119998812675476},{"id":"https://openalex.org/C127162648","wikidata":"https://www.wikidata.org/wiki/Q16858953","display_name":"Channel (broadcasting)","level":2,"score":0.4332999885082245},{"id":"https://openalex.org/C2776760102","wikidata":"https://www.wikidata.org/wiki/Q5139990","display_name":"Code (set theory)","level":3,"score":0.3361999988555908},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.32690000534057617},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.30149999260902405},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.30149999260902405},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.30090001225471497},{"id":"https://openalex.org/C179518139","wikidata":"https://www.wikidata.org/wiki/Q5140297","display_name":"Coding (social sciences)","level":2,"score":0.2994999885559082},{"id":"https://openalex.org/C28855332","wikidata":"https://www.wikidata.org/wiki/Q198099","display_name":"Quantization (signal processing)","level":2,"score":0.2994000017642975},{"id":"https://openalex.org/C45374587","wikidata":"https://www.wikidata.org/wiki/Q12525525","display_name":"Computation","level":2,"score":0.29269999265670776},{"id":"https://openalex.org/C99498987","wikidata":"https://www.wikidata.org/wiki/Q2210247","display_name":"Noise (video)","level":3,"score":0.2921000123023987},{"id":"https://openalex.org/C2781238097","wikidata":"https://www.wikidata.org/wiki/Q175026","display_name":"Object (grammar)","level":2,"score":0.27379998564720154},{"id":"https://openalex.org/C153258448","wikidata":"https://www.wikidata.org/wiki/Q1199743","display_name":"Gradient descent","level":3,"score":0.2612000107765198},{"id":"https://openalex.org/C105339364","wikidata":"https://www.wikidata.org/wiki/Q2297740","display_name":"Software deployment","level":2,"score":0.2603999972343445},{"id":"https://openalex.org/C155032097","wikidata":"https://www.wikidata.org/wiki/Q798503","display_name":"Backpropagation","level":3,"score":0.26030001044273376}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:doi:10.48550/arxiv.2602.15720","is_oa":true,"landing_page_url":null,"pdf_url":null,"source":{"id":"https://openalex.org/S4406922384","display_name":"Open MIND","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"repository"},"license":"publisher-specific-oa","license_id":"https://openalex.org/licenses/publisher-specific-oa","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Article"},{"id":"doi:10.48550/arxiv.2602.15720","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2602.15720","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"article"}],"best_oa_location":{"id":"pmh:doi:10.48550/arxiv.2602.15720","is_oa":true,"landing_page_url":null,"pdf_url":null,"source":{"id":"https://openalex.org/S4406922384","display_name":"Open MIND","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"repository"},"license":"publisher-specific-oa","license_id":"https://openalex.org/licenses/publisher-specific-oa","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Article"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Vision":[0],"Transformers":[1],"(ViTs)":[2],"have":[3,28],"achieved":[4],"remarkable":[5],"success":[6],"across":[7,111],"various":[8],"vision":[9],"tasks,":[10,152],"yet":[11],"their":[12],"deployment":[13],"is":[14],"often":[15],"hindered":[16],"by":[17],"prohibitive":[18],"computational":[19],"costs.":[20],"While":[21],"structured":[22,64],"weight":[23],"pruning":[24,65],"and":[25,39,118,129,163],"token":[26],"compression":[27,92],"emerged":[29],"as":[30],"promising":[31],"solutions,":[32],"they":[33],"suffer":[34],"from":[35],"prolonged":[36],"retraining":[37],"times":[38],"global":[40,96],"propagation":[41,97],"that":[42,90,122],"creates":[43],"optimization":[44],"challenges,":[45],"respectively.":[46],"We":[47,60],"propose":[48],"ToaSt,":[49],"a":[50],"decoupled":[51],"framework":[52],"applying":[53],"specialized":[54],"strategies":[55],"to":[56,66,74,150],"distinct":[57],"ViT":[58],"components.":[59],"apply":[61],"coupled":[62],"head-wise":[63],"Multi-Head":[67],"Self-Attention":[68],"modules,":[69],"leveraging":[70],"attention":[71],"operation":[72],"characteristics":[73],"enhance":[75],"robustness.":[76],"For":[77],"Feed-Forward":[78],"Networks":[79],"(over":[80],"60\\%":[81],"of":[82],"FLOPs),":[83],"we":[84],"introduce":[85],"Token":[86],"Channel":[87],"Selection":[88],"(TCS)":[89],"enhances":[91],"ratios":[93],"while":[94],"avoiding":[95],"issues.":[98],"Our":[99],"analysis":[100],"reveals":[101],"TCS":[102],"effectively":[103,149],"filters":[104],"redundant":[105],"noise":[106],"during":[107],"selection.":[108],"Extensive":[109],"evaluations":[110],"nine":[112],"diverse":[113],"models,":[114],"including":[115],"DeiT,":[116],"ViT-MAE,":[117],"Swin":[119],"Transformer,":[120],"demonstrate":[121],"ToaSt":[123,137,147],"achieves":[124,138],"superior":[125],"trade-offs":[126],"between":[127],"accuracy":[128,140],"efficiency,":[130],"consistently":[131],"outperforming":[132],"existing":[133],"baselines.":[134],"On":[135],"ViT-MAE-Huge,":[136],"88.52\\%":[139],"(+1.64":[141],"\\%)":[142],"with":[143],"39.4\\%":[144],"FLOPs":[145],"reduction.":[146],"transfers":[148],"downstream":[151],"achieving":[153],"52.2":[154],"versus":[155],"51.9":[156],"mAP":[157],"on":[158],"COCO":[159],"object":[160],"detection.":[161],"Code":[162],"models":[164],"will":[165],"be":[166],"released":[167],"upon":[168],"acceptance.":[169]},"counts_by_year":[],"updated_date":"2026-04-04T16:13:02.066488","created_date":"2026-02-19T00:00:00"}